Risks

The following risks have been identified regarding the success of our project.

1. Having difficulties for evaluating our overall result: we can compare two given curves, but it is impossible to look at every single pair.
   Since the human visual system is quite efficient at recognizing similar shapes, performing enough samples and manually evaluating their validity should be enough to make sure our algorithm works correctly.
2. Discovering that a heavy bias has been introduced by the OCR in some specific years, thus disturbing our whole distance metrics function.
   Such a situation would be handled by spending some time on cleaning the data.

Some trivial risks are omitted here, such as team member leaving the course, Le Temps denying us access to their data, lacking technical inspiration, and so on.